Gemini 3.5 Flash-Lite vs GPT-5.2-Codex
| Spec | Gemini 3.5 Flash-Lite | GPT-5.2-Codex |
|---|---|---|
| Respan Index | 29.9 (#92) | 30.1 (#90) |
| Blended price per 1M | $0.85 | $4.81 |
| Context window | 1.0M | 400K |
| Input / output per 1M | $0.3 / $2.5 | $1.75 / $14 |
| Max output | 66K | 128K |
| Input | text, vision, audio | text, vision |
| Released | 2026-07-21 | 2025-12-18 |
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | Gemini 3.5 Flash-Lite | GPT-5.2-Codex |
|---|---|---|
| Coding | 6.0 Gemini 3.5 Flash-Lite | 18.8 GPT-5.2-Codex |
| Agents | 44.0 Gemini 3.5 Flash-Lite | 22.2 GPT-5.2-Codex |
| Reasoning | 25.5 Gemini 3.5 Flash-Lite | 35.8 GPT-5.2-Codex |
| Knowledge & language | 32.0 Gemini 3.5 Flash-Lite | 34.4 GPT-5.2-Codex |
| Human preference | 52.1 Gemini 3.5 Flash-Lite | 49.4 GPT-5.2-Codex |
Benchmarks both report
GPT-5.2-Codex leads on 6 of the 8 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | Gemini 3.5 Flash-Lite | GPT-5.2-Codex |
|---|---|---|
| LMArena WebDev (LMArena) | 1432 | 1339 |
| LiveBench Agentic Coding (LiveBench) | 45.3% | 49.4% |
| LiveBench Coding (LiveBench) | 76.1% | 83.6% |
| LiveBench Data Analysis (LiveBench) | 53.2% | 78.2% |
| LiveBench Instruction Following (LiveBench) | 67.2% | 66.4% |
| LiveBench Language (LiveBench) | 71.8% | 73.7% |
| LiveBench Mathematics (LiveBench) | 73.7% | 88.8% |
| LiveBench Reasoning (LiveBench) | 60.2% | 77.7% |
FAQ
Is Gemini 3.5 Flash-Lite better than GPT-5.2-Codex?
GPT-5.2-Codex scores 30.1 on the Respan Index (#90), ahead of Gemini 3.5 Flash-Lite at 29.9 (#92). GPT-5.2-Codex leads on 6 of the 8 benchmarks both report.
Which is cheaper, Gemini 3.5 Flash-Lite or GPT-5.2-Codex?
Gemini 3.5 Flash-Lite lists at $0.3 input and $2.5 output per 1M tokens; GPT-5.2-Codex at $1.75 and $14. Gemini 3.5 Flash-Lite is 5.7x cheaper per token at list price.
Which has the larger context window?
Gemini 3.5 Flash-Lite accepts 1.0M tokens and GPT-5.2-Codex 400K tokens.
Compare other models
- vsGemini 3.5 Flash-Lite vs Gemini 3.1 Flash-Lite
- vsClaude Opus 5.5 vs GPT-5.2-Codex
- vsGemini 3.5 Flash-Lite vs Muse Glimmer
- vsGemini 3.5 Flash-Lite vs Qwen3.5 397B-A17B
- vsGemini 3.5 Flash-Lite vs Inkling-Small
- vsGPT-5.2-Codex vs MiMo-V2.5-Pro
- vsGPT-5.2-Codex vs Muse Glimmer
- vsGPT-5.2-Codex vs NVIDIA Nemotron 3.5 Super VL 120B-A12B NVFP4